feat: add MacOS support

[not-matthias]

No description provided.

[codspeed-hq]

Merging this PR will degrade performance by 54.04%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 9 improved benchmarks
❌ 3 regressed benchmarks
✅ 213 untouched benchmarks

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Memory	test_recursive_cached_fibo_30	1 B	29,968 B	-100%
❌	Simulation	test_recursive_fibo_10	139 µs	276.3 µs	-49.67%
❌	WallTime	switch 2	300 ns	336 ns	-10.71%
⚡	Memory	wait 1ms	16 B	10 B	+60%
⚡	Simulation	one	589.9 µs	402.7 µs	+46.5%
⚡	WallTime	test sync baz 10	384 ns	324 ns	+18.52%
⚡	WallTime	switch 1	324 ns	276 ns	+17.39%
⚡	WallTime	test sync baz 10	360 ns	312 ns	+15.38%
⚡	Simulation	wait 500ms	12.2 ms	10.7 ms	+14%
⚡	WallTime	iterative fibo 15	312 ns	276 ns	+13.04%
⚡	WallTime	iterative fibo 20	360 ns	324 ns	+11.11%
⚡	WallTime	test_iterative_fibo_10	396 ns	360 ns	+10%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing cod-2720-bump-instrument-hooks-in-codspeed-node-to-support-macos (f2b5162) with main (4dae798)}

[greptile-apps]

Greptile Summary

This PR adds macOS walltime support to the CodSpeed Node.js plugins by replacing the __codspeed_root_frame__ function-wrapping approach with timestamp-based marker pairs (currentTimestamp + addMarker), and introduces a dedicated codspeed-walltime-macos CI job that runs a macOS-only vitest benchmark.

• Measurement strategy: packages/vitest-plugin/src/walltime/index.ts now captures currentTimestamp() before/after originalRun and emits MARKER_TYPE_BENCHMARK_START/END markers per task; packages/tinybench-plugin/src/shared.ts adds an equivalent sendRunMarkers that emits a single pair covering the entire batch run (walltime mode only).
• CI restructuring: The old walltime-macos-test job is removed from ci.yml and replaced by codspeed-walltime-macos in codspeed.yml, which targets only macos.bench.ts to avoid duplicate uploads of the Linux benchmarks.
• Submodule bump: packages/core/src/native_core/instruments/hooks is updated to expose the new currentTimestamp and addMarker native APIs.

Confidence Score: 4/5

Safe to merge once the staging upload URL is resolved; the measurement logic changes are well-structured.

The macOS CI job hardcodes the staging upload URL, meaning every macOS benchmark upload silently goes to staging rather than production. No macOS walltime data will ever appear in the production dashboard until this is corrected. All other changes look correct.

.github/workflows/codspeed.yml — the upload-url field in codspeed-walltime-macos points at the staging API.

Important Files Changed

Filename	Overview
.github/workflows/codspeed.yml	Adds codspeed-walltime-macos job, but hardcodes the staging upload URL (api.staging.preview.codspeed.io) which will send all macOS benchmark data to the staging environment rather than production.
.github/workflows/ci.yml	Removes the old walltime-macos-test CI job that was replaced by codspeed-walltime-macos in codspeed.yml.
packages/vitest-plugin/src/walltime/index.ts	Replaces the __codspeed_root_frame__ function wrapper approach with timestamp-based markers (currentTimestamp + addMarker) per task, aligning the vitest plugin with the new marker-based walltime measurement strategy.
packages/tinybench-plugin/src/shared.ts	Adds sendRunMarkers that emits a single BENCHMARK_START/BENCHMARK_END marker pair (in walltime mode only) covering the entire batch run, gated by getInstrumentMode() !== "walltime".
packages/vitest-plugin/benches/macos.bench.ts	New macOS-only benchmark file that skips on non-Darwin platforms; run exclusively by the codspeed-walltime-macos CI job to avoid duplicate uploads.
packages/tinybench-plugin/tests/index.integ.test.ts	Adds mocks for the two new InstrumentHooks methods (currentTimestamp, addMarker) so the integration test compiles cleanly with the updated shared.ts.
packages/core/src/native_core/instruments/hooks	Submodule pointer bumped to b9ddb5b to pick up the new native currentTimestamp and addMarker APIs required by this PR.

Sequence Diagram

sequenceDiagram
    participant VW as WalltimeRunner
    participant TB as WalltimeBenchRunner
    participant IH as InstrumentHooks
    participant CS as CodSpeed API

    Note over VW: Per-task flow (vitest)
    VW->>IH: startBenchmark()
    VW->>IH: currentTimestamp() runStart
    VW->>IH: originalRun.call(this)
    VW->>IH: currentTimestamp() runEnd
    VW->>IH: stopBenchmark()
    VW->>IH: addMarker(BENCHMARK_START, runStart)
    VW->>IH: addMarker(BENCHMARK_END, runEnd)
    VW->>IH: setExecutedBenchmark(pid, uri)

    Note over TB: Batch flow (tinybench)
    TB->>IH: currentTimestamp() batchStart
    loop each task
        TB->>IH: startBenchmark()
        TB->>IH: task.run()
        TB->>IH: stopBenchmark()
        TB->>IH: setExecutedBenchmark(pid, uri)
    end
    TB->>IH: currentTimestamp() batchEnd
    TB->>IH: addMarker(BENCHMARK_START, batchStart)
    TB->>IH: addMarker(BENCHMARK_END, batchEnd)
    IH-->>CS: upload results (walltime)

Loading

_{Reviews (13): Last reviewed commit: "chore: run on staging" | Re-trigger Greptile}

[greptile-apps]

Mutable branch reference for runner-version

runner-version: branch:sip-resign-exec-redirect points to a live branch that can be force-pushed or rebased at any time. If the branch tip changes between two benchmark runs, the profiling behaviour changes silently, making measurement results non-comparable across commits. Consider pinning to a specific commit SHA or tag once the feature branch is stable, e.g. runner-version: sha:<commit>.

[codspeed-hq]

Unable to generate the flame graphs

The performance report has correctly been generated, but there was an internal error while generating the flame graphs for this run. We're working on fixing the issue. Feel free to contact us on Discord or at support@codspeed.io if the issue persists.

[GuillaumeLagrange]

olgtm

[GuillaumeLagrange]

dont forget to drop this

[greptile-apps]

Staging upload URL will suppress macOS results in production

upload-url: https://api.staging.preview.codspeed.io/upload routes every upload from this job to the staging environment. macOS benchmark results will never appear in the production CodSpeed dashboard for any commit — the data lands in staging silently. If the intent was to test the macOS runner before going to production, this URL should be removed (or replaced with the production endpoint) once validation is complete.

[avalanche-staging]

Congrats! CodSpeed is installed 🎉

🆕 1 new benchmark was detected.

You will start to see performance impacts in the reports once the benchmarks are run from your default branch.

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

Detected benchmark

• fibo darwin (WallTime): 8.3 ms

---